Comparing Cobb–Douglas and translog stochastic frontier models for estimating technical efficiency in rice farming in Northwestern Nigeria

This study explores the technical efficiency of rice farmers in northwestern Nigeria by comparing the Cobb–Douglas and translog production functions. To address the gap in understanding the most suitable model, a multistage sampling technique was used to select 370 farmers, and the data were analysed via the stochastic frontier model. The findings indicate that the translog model, which is superior to the Cobb–Douglas model, resulted in increasing returns to scale, with significant effects on farm size, seeds, and fertilizer. This study revealed few significant input interactions, notably between labour-chemicals, seed-fertilizers, and seed-chemicals, highlighting the importance of complementary input use in optimizing rice production efficiency. Cobb–Douglas revealed decreasing returns with significant changes in farm size, labour, seeds, and fertilizer. The Harman single-factor test revealed no significant common method bias in the data, confirming the validity of the findings and enhancing the reliability of the estimates. Sex, education, and poverty status positively influence efficiency. The negative factors included land rights, distance to market, and livestock size. Overall, the translog production function was recommended for accurately estimating the technical efficiency of rice production, emphasizing the need for appropriate model selection on the basis of statistical properties. To increase the technical efficiency and productivity among rice farmers in northwestern Nigeria, agricultural policies should prioritize access to quality seeds, fertilizer, and optimized farm size management, as these inputs significantly influence efficiency under the preferred translog production function.